New tool twice as accurate at predicting antibody resistance among U.S. children with Kawasaki disease

Dec. 5, 2022
UB-developed scoring system lays groundwork for North America’s first test to predict therapeutic resistance for the rare disease.

A new tool under development by University at Buffalo researchers could one day help clinicians better predict resistance to immunoglobulin therapy among children with Kawasaki disease in the United States.

The new tool was described in a study published earlier this year in the Journal of Pediatric Pharmacology and Therapeutics.

Compared to the Kobayashi score – the most widely used method in Japan for predicting resistance to the antibody, intravenous immunoglobulin (IVIG), in Kawasaki disease – the UB-developed tool is twice as accurate for Western New York children, achieving a sensitivity (detection) rate of 54%.

The UB study analyzed 208 cases of children hospitalized with Kawasaki disease at a Buffalo hospital between 2000-15. The data were divided into two groups: IVIG susceptible or resistant.

Four variables that differed between the groups were identified and used to create a new prediction score. The variables included: white blood cell count, neutrophil percentage, age, and serum albumin concentration. Each variable is assigned one point and a score of three or greater denotes high risk for IVIG resistance.

To ensure the new scoring system is applicable across geographical locations, the accuracy of the tool was tested on Kawasaki disease patient data obtained from Western New York, Boston and China.

The new tool had a sensitivity rate of 54% for the Western New York data set, while the rate for the Kobayashi score was 26%. For the Boston data set, the new score had a sensitivity of 40%; far higher than the 0% rate for the Kobayashi score. The sensitivity rate for the new score was the lowest among the China data set, at 27%, whereas the Kobayashi score had a sensitivity rate of 36%.

University at Buffalo release